As the holiday rush kicks into gear, product teams everywhere are racing to get thousands of product detail pages (PDPs) live before Black Friday. Generative AI makes that race feel winnable; a magic tool can write a description in seconds, and a whole catalog in an afternoon.

But speed has a side effect. Shoppers are scrolling past descriptions that all sound the same, and AI engines are skipping products that say a lot without saying anything. The result is a growing pile of content we now call AI slop, and it is quietly costing brands sales.

The good news? Avoiding it does not mean abandoning AI. It means pointing AI at the right foundation: your product data.

What is AI Slop?

Merriam-Webster made “slop” its 2025 Word of the Year, defining it as low-quality digital content, typically churned out in volume by artificial intelligence.

On a product detail page, AI slop is copy that looks finished but carries no real information. It is written to fill a field, not to help a buyer decide. You have probably seen it:

• Empty superlatives: “Elevate your everyday with timeless style and unmatched quality.”
• Invented or vague specs: “Long-lasting battery” with no hours listed, or worse, a figure that does not match the spec sheet.
• Copy-paste sameness: Ten variants of the same jacket, each with an identical paragraph that never mentions the fabric weight or fit.
• Answers to nobody’s question: Three sentences about “versatility” and zero about whether it fits a carry-on bin.

Why AI Slop Costs You Twice

It loses the human shopper. Shoppers do their homework. 89% of consumers research products before buying, and 69% have returned a deal-day purchase. When the details aren’t there, they leave: 70% of shoppers said they would buy a different product due to missing information, and 43% had returned something in the past year because pre-purchase information was wrong.

It loses the AI engine. Discovery is moving off your website. McKinsey projects that $750 billion in US revenue will flow through AI-powered search by 2028. AI engines don’t reward fluent filler. They look for structured, specific, verifiable facts they can compare and cite. Slop gives them nothing to extract, so your product lands on what we call the Invisible Shelf: online, but never recommended.

What a Good PDP Looks Like

Take a women’s waterproof rain jacket, a category where sizing and fit questions drive a huge share of returns. Here’s the same product, described two ways:

A great PDP does two jobs at once. It tells a human shopper a persuasive, accurate story, and it gives AI engines clean, structured facts (backed by Product Schema) that they can parse, compare and recommend with confidence.

6 Ways to Keep Slop Off Your PDPs 

1. Feed AI facts, not a product name 

AI only writes slop when it has to guess. Before generating copy, make sure the product record holds the attributes that matter: dimensions, materials, compatibility, care, certifications. Rule of thumb: if a fact isn’t in your product data, it shouldn’t appear in the description.

2. Run the “swap test” 

Paste your description onto a competitor’s product page. If it still reads as true, it’s slop. Rewrite until every sentence could only describe your product.

3. Replace adjectives with numbers

“Lightweight” becomes “1.2 kg.” “Fast charging” becomes “0 to 80% in 35 minutes.” Search your catalog for words like premium, innovative and seamless, then swap each for a measurable fact.

4. Answer the questions buyers actually ask 

Pull the top questions from customer service tickets, reviews, and on-site search. Add short FAQs to the PDP, such as “Is it machine washable?” These pre-packaged answers are exactly what AI engines lift into their responses; plus you know you’ll be speaking the same language as your customers since you used their feedback to feed your product information.

5. Map products to real use cases 

Add structured attributes for occasions, environments, and personas: “apartment-friendly,” “gift under £100,” “rated for outdoor use.” Context is what connects your product to a conversational query.

6. Keep a human in the loop and one source of truth 

Set brand voice rules and review workflows so a person approves AI-generated copy before it goes live. Then syndicate that approved version everywhere, so your website, marketplaces and Google Merchant Center never contradict each other.

Get Off the Invisible Shelf Before the Holidays

Despite the name, AI isn’t necessarily the villain in this situation. Often, it can be a great writing partner when you give it something real to work with. Slop happens when we ask it to fill in the blanks, and it happily obliges.

So before you hit “generate” on another few hundred descriptions this holiday season, take a quick look at what you’re feeding it. Pick a handful of your best sellers, run the swap test, and trade a few adjectives for facts. You’ll be surprised how quickly the copy starts to sound like your product again.

Your shoppers will notice. And increasingly, so will the AI engines deciding which products to recommend.

As AI becomes more deeply embedded in product operations, governance is moving quickly from a theoretical concern to a practical one. 

Product teams are starting to build agents that recommend attributes, generate content, flag compliance issues, prepare localized versions, and influence what ultimately reaches customers across different channels.

Most organizations recognize that these capabilities need boundaries. They create policies that define which tools may be used, which data sources are approved, when human review is required, and which decisions should never be automated. 

That is an essential starting point, but it does not answer every question that arises once AI becomes part of a live product workflow.

A policy can tell you what should have happened. It cannot necessarily show you what actually happened.

That is where decision lineage comes in. Decision lineage creates a connected record of how source information became an AI recommendation, how that recommendation was reviewed or changed, and how the final decision reached a product record or live channel. It turns governance from a set of written expectations into evidence that can be inspected, explained, and improved.

As product teams introduce more agents and automate more of the product lifecycle, that evidence becomes increasingly important. Without it, even a well-written governance policy can leave teams piecing together decisions across approval tools, system logs, emails, and spreadsheets. With it, they can trace a product claim back to its source, understand why a decision was made, correct errors more quickly, and use recurring patterns to strengthen future rules and workflows.

An effective AI governance approach therefore needs both: a policy that defines how AI should operate, and decision lineage that shows how it did.

What Is an AI Governance Policy?

An AI governance policy is the rulebook for how your organization selects, builds, deploys, and uses AI.

For product data teams, it should define:

• Which AI use cases and data sources are approved
• Who owns each agent and the decisions it supports
• Which actions may be automated and which require human review
• How quality, security, privacy, compliance, and operational risks are managed
• How incidents, exceptions, and system changes are documented and escalated

A useful policy should guide everyday decisions without trying to predict every future AI use case. NIST’s AI Risk Management Framework treats governance as an ongoing part of designing, deploying, using, and evaluating AI systems, not as a document that is completed once before launch.

The policy establishes boundaries. It does not, on its own, create evidence that those boundaries were followed; that’s where decision lineage comes into play.

What Is Decision Lineage?

Decision lineage is the connected record of how an AI-assisted recommendation moved from source information to a business outcome.

For product data, that record should show:

• The source and version used
• The agent, model, prompt, mapping, or business rule applied
• The recommendation and supporting evidence produced
• Any confidence, quality, or risk signal
• Who approved, edited, overrode, or rejected the recommendation
• The resulting change to the product record
• The downstream agents, workflows, and channels affected

Decision lineage is more than a model log, an approval timestamp, or product version history. Those are individual pieces of the record, and lineage connects them.

That becomes essential when several agents act on the same product. A Supplier Data Agent might update a specification while an Enrichment Agent uses that value to complete related attributes and a Content Agent then turns those attributes into customer-facing copy.

If the original specification is later challenged, your team needs one connected history, not three separate logs held by different systems.

This is what we like to call unified observability: a shared history connecting source evidence, agent actions, human reviews, and product changes.

Why an AI Governance Policy Needs Decision Lineage

A policy can state that regulated or commercially sensitive claims require human review. Decision lineage shows whether that review happened before publication, what the reviewer considered, and whether a later change invalidated the original approval.

This distinction is reflected in established AI-governance guidance.

NIST recommends maintaining histories and audit logs, documenting the degree of human oversight, tracking overrides and reported errors, and recording accountable “go” and “no-go” decisions.

The OECD’s AI Principles similarly emphasize giving stakeholders meaningful information about the data, inputs, factors, processes, or logic behind an AI outcome. Its guidance describes traceability as a way to support later analysis, understand outcomes, prevent repeated mistakes, and improve trustworthiness.

For high-risk AI systems, the EU AI Act includes requirements for automatic event logging and effective human oversight. Not every product-data agent will fall within that classification, but the direction is useful: responsible governance increasingly requires evidence of system behavior, not simply written statements of intent.

How to Build Decision Lineage Into Your Governance Policy

1. Start With Decisions, Not Tools

Create an inventory of the decisions each agent makes or influences.

“We use generative AI for product content” is too broad. “The Enrichment Agent recommends missing dimensions from approved supplier documents” is specific enough to govern.

For each decision, record the product families affected, the potential consequence of an error, and whether the action can be reversed. This helps you distinguish routine automation from decisions that need stronger controls.

2. Define the Boundaries of Automation

Specify what each agent may complete automatically, what it may recommend for approval, and what must remain human-led.

Use practical review triggers rather than a general instruction to “keep a human in the loop.” Human review might be required when:

• Confidence falls below an agreed threshold
• Two authoritative sources conflict
• A regulated, technical, or sustainability claim is introduced
• A change would affect a large number of SKUs
• The agent is acting outside its usual product family or market

Assign the reviewer, explain what evidence they need, and define how they can approve, override, correct, or roll back the action.

3. Establish Source Authority

Your policy should identify which sources are authoritative for each category of product information.

A current manufacturer datasheet may outrank a legacy spreadsheet for technical dimensions. An approved certification may be mandatory before a sustainability claim can be published. A retailer specification may control formatting without becoming the source of truth for the underlying product value.

The policy should also explain what happens when sources disagree. An agent should not quietly select whichever value appears most convenient or most frequently.

4. Define the Minimum Lineage Record

Require every material AI-assisted decision to capture a consistent set of information:

• Source owner, version, timestamp, and authority level
• Agent, model, prompt, mapping, and rule versions
• Input, recommendation, evidence, and confidence signal
• Reviewer, final decision, edits, and reason for any override
• Product record or attribute changed
• Downstream workflows and channels affected

The amount of detail can vary according to risk. An image-tag suggestion does not need the same record as a safety, technical, or environmental claim.

5. Connect Lineage Across Agents

Give each material change a shared identifier that follows it through enrichment, validation, content creation, localization, and activation.

This makes it possible to answer a practical question: if the original source value changes tomorrow, which descriptions, translations, product records, and published destinations need to be reviewed again?

Isolated logs tell you what each tool did. Connected lineage tells you what happened to the product.

6. Run Reconstruction Drills

Do not wait for a retailer, customer, auditor, or regulator to test your records.

Periodically select a published claim or attribute and work backwards. Can your team identify:

• The original source
• The automated actions taken
• The rules applied
• The people who reviewed the recommendation
• Any edits or overrides
• Every channel that received the result

Then test whether the information can be stopped, corrected, and republished.

Track how long reconstruction takes, where evidence is missing, how often reviewers override agents, and which corrections repeat. Those patterns should feed back into your data model, prompts, mappings, business rules, and review thresholds.

From Written Rules to Operational Evidence

As AI becomes part of everyday product operations, governance cannot remain a document that teams consult only during procurement, implementation, or compliance reviews. It needs to be reflected in the way product information is sourced, recommendations are made, approvals are recorded, and changes move through the product lifecycle.

A well-designed governance policy provides the necessary structure. It establishes ownership, defines acceptable uses, sets boundaries around automation, and clarifies when human judgment is required. Decision lineage carries those principles into practice by preserving the evidence behind each material decision: the source an agent relied on, the rules it applied, the recommendation it produced, the person who reviewed it, and the product records or channels affected.

As more agents begin contributing to product information, the ability to generate an answer will be only one part of responsible AI. Teams will also need to explain where that answer came from, how it was evaluated, why it was accepted, and what happened next. That is what turns an AI governance policy from a statement of intent into a system people can understand, trust, and continuously improve.

In Part 1 of this series, we explored how DwyerOmega rebuilt its product data pipeline using Akeneo MCP Server, creating a stable, self-updating connection between Akeneo and the systems its teams use every day.

With Akeneo remaining the governed source of product information, the next question was what DwyerOmega could do once that trusted data became easier to access and activate.

The answer is already taking shape across the business.

Bringing product data quality into Power BI

With the MCP-powered pipeline in place, product management and marketing teams get a familiar, filterable view of catalog quality in Power BI. The dashboards show whether products meet their targets across areas such as:

• Hero and variant-level images
• 360-degree product spins
• Image alt text
• Short and long descriptions
• Unique SKU-level descriptions
• Accessories and product relationships
• CAD files
• Upsell and cross-sell assignments
• SEO titles and metadata

Users can filter by product manager or product model, see where current content does not meet its target, and export a focused list of products that need attention. The dashboard is intentionally read-only.

When asked whether edits in Power BI push back into the PIM, Cyrus Grimes, Director of Product Content Management at DwyerOmega, was direct: “No, this is just a report.” Updates are still made in Akeneo or through coordinated loading documents handled by trained PIM users.

That combination gives teams convenient access without weakening control. In one example, a manager used the dashboard to show a product owner exactly where content and image targets were misaligned.

The product owner could then defend the targets or improve the data, working from a clear list of affected products rather than a broad request to fix completeness.

Moving from reactive fixes to proactive quality management

Before this visibility existed, inconsistencies could surface through customer complaints, manual website checks, or one-off investigations. Power BI gives DwyerOmega a more proactive way to find missing or misaligned information before it becomes a downstream problem.

Because the company is in an active acquisition phase rather than a maintenance phase, data quality work has historically been easy to push to the back burner. The dashboard makes that work visible and assignable. Cyrus plans to extend it as acquisitions become a smaller part of daily operations, including new status views and additional objects.

Two further ideas are in early exploration rather than production: a data hygiene or “policing” view that would compare records across the source system, SAP Commerce Cloud, and Akeneo to surface synchronization gaps, and an AI-assisted workflow that reads product data issues reported in a team chat channel, checks Akeneo, and proposes corrections through MCP with human approval retained.

The operating model is clear: automate repeatable discovery and preparation, then keep people in control of consequential updates.

Turning technical documentation into stronger product descriptions

DwyerOmega is also using a multi-phase, AI-assisted workflow built in Claude to improve product descriptions. Many technical products have capabilities documented in specification sheets and manuals but represented online by only the small set of popular variants loaded as PIM variants. Older descriptions were often thin.

“We had some product descriptions that were pretty weak. Literally just a paragraph, and kind of just marketing fluff.”

The workflow uses Akeneo product records to locate the relevant manuals and specification sheets. Local optical character recognition, run through a local language model, extracts the technical content, followed by quality control. The process then builds structured, capability-led descriptions and merges them with useful elements of the existing copy before product managers review and load them back into Akeneo.

Preserving the existing content matters. Cyrus reports that the approach has helped keep organic search traffic stable and growing despite broader industry declines.

Accelerating the onboarding of acquired product lines

The most immediately valuable use case is the acquisition workflow, and Cyrus names it as the one with the greatest potential.

Newly acquired businesses do not organize or sell products the way DwyerOmega’s established ecommerce catalog does. Some products are highly technical, expensive, or configurable across an impractical number of permutations.

“Not everything is easy to add to your cart and sell there. They might be extremely expensive or very technical, and we do not want the customer to just spend $10,000 and be like, ‘I got the wrong thing’.”

To support these products, DwyerOmega created a sales-assisted product family where the call to action is contacting a salesperson rather than adding to cart, while the product information is still fully managed and published.

MCP-powered workflows chain several systems together. A web scrape of an acquired company’s product series loads supporting assets into SharePoint, mirrors images into AWS for web delivery, creates the asset in Akeneo with the URL populated automatically, and builds the product series with a placeholder variant and basic information.

The workflow also reduces a common acquisition tradeoff. DwyerOmega does not have to wait for every commerce decision to be finalized before making newly acquired product information useful to customers and sales teams.

The impact: Akeneo data works across the business

Across these use cases, DwyerOmega is building an operating model in which Akeneo remains the authority for product information and MCP reduces the friction between that source and the systems where work happens.

The practical benefits are already visible:

1. Faster acquisition integration. Website content and assets become governed product structures in about 40 minutes per series, with 82 series loaded in the first few weeks.
2. A more resilient data foundation. Automated delta updates replaced a less stable extraction pipeline and supported growing multilingual volume.
3. More accessible product intelligence. Product, marketing, SEO, and sales operations teams work in familiar Power BI views instead of filing ad hoc requests.
4. Clearer priorities for improvement. Teams move from broad completeness goals to targeted, exportable lists of products that need attention.
5. Stronger content operations. Technical documents inform structured product descriptions through a repeatable, reviewable workflow.
6. Human expertise stays in the loop. Akeneo governance and human review protect quality while MCP handles discovery and preparation.

The value of governed product data grows when people and systems can use it without introducing another layer of inconsistency. For DwyerOmega, Akeneo remains the trusted product foundation, while MCP makes that information easier to access and put to work across the business. What started as a better way to connect systems is becoming a more flexible operating model for product information.

Black Friday shopping used to be pretty straightforward. 

You made a list, hunted down a few discounts, and tried not to get distracted by a suspiciously cheap television you absolutely didn’t need.

Now? You’ve got 17 browser tabs open, three competing discount codes, an AI assistant comparing product reviews, and a group chat debating whether that “limited-time deal” was actually cheaper last Tuesday.

Welcome to holiday shopping in 2026.

And while the shopping experience might look a little different these days, one thing hasn’t changed: there’s a whole lot of money up for grabs.

According to Adobe’s 2026 Holiday Shopping Forecast, U.S. online holiday spending is expected to reach a record $275.1 billion in November and December, with Cyber Monday alone projected to generate $15.1 billion.

But with shoppers becoming savvier, AI playing a bigger role in purchasing decisions, and the lines between online and offline shopping blurring, a flashy discount isn’t necessarily enough to seal the deal anymore.

So, before the holiday madness officially begins, let’s unwrap six trends shaping Black Friday and Cyber Monday 2026, and what brands can do to make sure they’re on shoppers’ nice lists.

1. Black Friday Is No Longer Just a Day (or Even a Weekend)

Remember when holiday shopping officially kicked off the morning after Thanksgiving? Those days are long gone.

According to the National Retail Federation (NRF), 43% of holiday shoppers plan to begin browsing and buying before November. However, it seems like we haven’t completely abandoned tradition altogether; 64% say they’re still likely to wait until Thanksgiving weekend to complete most of their shopping.

In other words, the holiday season now has multiple starting lines.

How brands can prepare: Think beyond a single promotional weekend, and start earlier than you think you need to (like, yesterday). Refresh holiday product content early, coordinate promotional calendars, and ensure inventory information stays accurate as demand builds.

Pay attention to what shoppers are searching for and comparing before they purchase. Those early signals can help brands refine their product experiences ahead of the biggest shopping days.

2. AI Assistants Are Becoming the New Holiday Shopping Guides

What happens when a shopper asks an AI assistant for the best holiday gifts under $100, and your product doesn’t make the shortlist?

That’s becoming a very real concern for brands.

According to Adobe’s 2026 Holiday Shopping Forecast, traffic from AI-powered tools to U.S. retail websites is expected to increase by 130% this holiday season compared to last year. Meanwhile, Gartner research found that 34% of U.S. consumers are very or extremely willing to use generative AI for holiday gift recommendations.

For brands, getting your product to show up in traditional search engines is hard enough, but ensuring AI systems can understand what makes their products relevant, compare them accurately against alternatives, and confidently recommend them for specific needs is the real challenge these days.

How brands can prepare: Invest in Answer Engine Optimization (AEO) by making product information structured, comprehensive, and accessible to AI-powered discovery tools. Go beyond basic descriptions to include detailed attributes, use cases, compatibility information, and verified product claims. After all, even the best Black Friday deal won’t matter if your product never makes it into the conversation.

3. Shoppers Want Proof of Value, Not Just Bigger Discounts

Everyone loves a good deal, but today’s shoppers are becoming increasingly skeptical about what actually constitutes one.

Akeneo’s latest research found that 79% of consumers have delayed a purchase because they expected the price to drop, while 77% have noticed price differences for the same product across retailers or platforms.

Meanwhile, only 32% completely or mostly trust retailers to offer fair or competitive prices.

The takeaway? A discount might attract attention, but it doesn’t automatically earn trust.

How brands can prepare: Bring product data and pricing intelligence closer together. Ensure channel-specific prices, promotional terms, and product variations are accurately represented.

Just as importantly, communicate what makes a product worth purchasing, whether that’s durability, quality, features, warranties, or convenience.

4. The Holiday Shopping Journey Is Happening Everywhere

A shopper discovers a product on TikTok, researches it through an AI assistant, compares prices on a marketplace, and eventually purchases it in-store.

Welcome to modern holiday shopping.

According to Salesforce’s 2026 holiday predictions, 28% of Gen Z shoppers plan to shop through social apps this holiday season. The same study also found that 79% of in-store shoppers use their phones while browsing.

The boundaries between online and offline shopping are becoming increasingly blurred.

How brands can prepare: Deliver consistent, accurate product information across every channel, from eCommerce sites and marketplaces to social platforms and physical stores. This includes everything from product descriptions and specifications to pricing, availability, and promotional messaging.

And consistency isn’t just about creating a seamless experience for shoppers; it’s becoming essential for AI discoverability. AI agents rely on trustworthy, authoritative product information to confidently recommend products. When they encounter conflicting specifications, inconsistent pricing, or contradictory descriptions across channels, it becomes harder to determine which information is reliable, making them less likely to cite or recommend your product.

After all, if an AI assistant can’t confidently figure out what you’re selling, why would it recommend it to someone else?

5. Product Information Will Make or Break Customer Confidence

 A great Black Friday discount might convince someone to make a purchase, but what happens when the product arrives and doesn’t quite live up to expectations?

Maybe that sweater runs two sizes too small, the headphones aren’t compatible with their device, or the product looks nothing like the photos online.

Suddenly, that fantastic holiday deal has turned into a frustrating return.

And it’s a surprisingly common problem. Last year, 69% of consumers returned a purchase made during a major deal event, with poor product quality, misleading imagery, and inaccurate descriptions among the reasons for dissatisfaction.

For brands, these returns don’t just eat into holiday profits; they can also undermine customer trust and make repeat purchases less likely.

How brands can prepare: Focus on helping customers make informed decisions before they buy. Provide accurate sizing guides, detailed specifications, realistic product imagery, compatibility information, and clear descriptions of what’s included in the purchase.

Encourage authentic customer reviews and address frequently asked questions directly on product pages. The more confidently customers can assess whether a product meets their needs, the less likely they are to experience buyer’s remorse.

6. AI-Powered Operations Will Separate the Prepared From the Overwhelmed

Behind every successful holiday shopping experience is an enormous amount of coordination.

Supplier updates, promotional changes, product enrichment, localization, marketplace requirements, and inventory adjustments all need to happen quickly. AI agents can increasingly help teams identify missing information, recommend improvements, and accelerate repetitive workflows.

But speed without oversight creates its own problems.

That’s why Akeneo’s agentic Product Cloud emphasizes coordinated AI capabilities alongside human expertise and governance.

How brands can prepare: Establish clear product readiness standards, approval processes, and ownership of critical product information. Use customer questions, search behavior, channel feedback, and returns data to identify opportunities for continuous improvement.

Instead of treating product information as something to finalize before Black Friday, businesses should approach it as a living asset that evolves throughout the shopping season.

The Holiday Rush Is Complicated Enough. Your Product Experience Shouldn’t Be.

Between the endless browser tabs, mysterious discounts, and AI assistants weighing in on everything from winter coats to kitchen appliances, holiday shoppers have plenty to think about. The last thing they need is conflicting product information, inaccurate pricing, or a purchase that doesn’t live up to expectations.

For brands, Black Friday and Cyber Monday in 2026 and beyond success should be less about offering the biggest discounts or running the flashiest campaigns, and more about making products easy to discover, compare, trust, and purchase, whether someone’s browsing your website, scrolling through social media, or asking an AI assistant for recommendations.

The good news? You don’t need a crystal ball to prepare for the future of commerce. With accurate, comprehensive product information, connected channels, and the right AI-powered tools, you can deliver better shopping experiences long after the holiday decorations come down.

After all, the best holiday shopping experience is the one that leaves your customers confident they made the right purchase, and ready for some holiday cheer.

Remember when the biggest innovation in online shopping was being able to buy something without leaving your couch?

We’ve come a long way since then.

Today, you can upload a photo of a pair of shoes to find something similar, ask an AI assistant to compare the best options for your budget, and get a personalized recommendation before you’ve even opened a retailer’s website.

And if 2026 taught us anything, it’s that artificial intelligence isn’t just changing how we shop. It’s changing what shopping actually looks like.

As we head into 2027, we’re moving into an era where search engines are giving way to conversations, product pages are becoming more contextual, and the line between human shoppers and AI-powered buying agents is starting to blur.

But don’t worry, we’re not predicting a world where robots are wandering around shopping malls with credit cards. At least, not yet.

Instead, we’re looking at the technologies, consumer behaviors, and industry developments already reshaping commerce to understand what comes next.

So, grab your metaphorical crystal ball. Here are seven eCommerce trends that we believe will define 2027.

1. Agentic Commerce Moves From Experimentation to Execution

Last year, we talked about the first steps toward agentic commerce: a future where AI assistants could help consumers move from discovering a product to actually purchasing it.

In 2027, that future is becoming much more tangible.

Think about the difference between asking an AI assistant to recommend a suitcase and asking it to find a carry-on that meets your airline’s size restrictions, fits your budget, arrives before Friday, and can be purchased from a retailer you trust.

The first requires a recommendation. The second requires an AI agent that can understand multiple constraints, compare real-time information, and potentially take action on your behalf.

And consumers are increasingly open to the idea. According to DHL’s 2026 eCommerce Trends Report, 29% of shoppers say they’d be happy to let AI make shopping decisions or purchases for them within the next five years.

The technology is evolving to support that shift, too. Google’s Universal Commerce Protocol is helping establish common standards for agent-driven shopping, while emerging checkout integrations are making it possible for purchases to take place within AI-powered experiences.

Of course, this doesn’t mean traditional eCommerce websites will disappear overnight. Shoppers will still want control over their purchases, and retailers will remain central to payment, fulfillment, and customer relationships.

But brands will increasingly need to think beyond optimizing their websites for human visitors.

In 2027, your next customer might be an AI agent shopping on behalf of a human. And that agent will need accurate, structured product information to decide whether your product makes the cut.

2. Product Discovery Shifts From Keywords to Intent

For years, eCommerce discovery has largely revolved around keywords.

Want a new winter coat? Search for “women’s waterproof winter coat,” scroll through a few pages of results, and hope something catches your eye.

But what if you could simply explain what you needed?

“I’m looking for a warm, waterproof jacket for walking my dog in the rain, but I don’t want anything too bulky, and I’d like to spend less than $150.”

That’s the kind of nuanced request AI-powered discovery is increasingly designed to understand.

And shoppers are already embracing it.

According to Adobe’s September 2026 Digital Insights research, referral traffic from AI platforms to UK retail websites increased by a staggering 1,439% between October 2024 and August 2026. By August 2026, AI-referred traffic was converting at a rate 20% higher than non-AI referred traffic.

Meanwhile, Salesforce’s 2026 State of Commerce research found that shoppers increasingly begin their purchasing journey with conversational AI, with the use of agentic search as a first step growing 200% year over year.

This shift introduces a new challenge for brands: how do you make sure your products are discovered when customers aren’t searching for your brand name, specific product category, or even a conventional keyword?

Enter Answer Engine Optimization (AEO), also called Generative Engine Optimization (GEO).

Rather than focusing exclusively on rankings in traditional search results, businesses need to make their product information understandable, relevant, and trustworthy to AI-powered discovery systems.

That means detailed attributes, structured specifications, clear product relationships, accurate availability, and content that explains what a product is actually good for.

Because when an AI assistant is trying to find the perfect winter coat, it needs more than a catchy product description and a few flattering photos. It needs to understand why that particular coat is the right answer.

3. Every Shopper Becomes an Expert Shopper

Remember when buying a complicated product meant spending hours researching specifications, reading reviews, watching comparison videos, and hoping you understood enough to make the right decision?

AI is starting to change that equation.

In 2027, expect consumers to become increasingly sophisticated shoppers, not necessarily because they know more about the products themselves, but because AI can bring expert-level knowledge into the buying process.

This becomes particularly interesting in B2B eCommerce, where purchasing decisions often involve technical specifications, compatibility requirements, certifications, and complex product configurations.

An AI assistant could help a buyer identify the right industrial replacement component based on operating conditions, existing equipment, and required safety certifications.

But here’s where things get interesting for brands and retailers: as shoppers become more informed, product experiences will need to become more contextual.

Rather than presenting every customer with the exact same product page, commerce experiences could increasingly assemble information around what a particular shopper is trying to accomplish, whether that’s highlighting a product’s dimensions, sustainability credentials, compatibility, or total cost.

The result is a potential shift away from one-size-fits-all product experiences toward shopping journeys built around individual intent.

However, there’s a catch: AI can only provide expert guidance when it has access to expert-level product information.

In a world where every customer can ask increasingly sophisticated questions, incomplete product descriptions and missing attributes will become harder to hide.

4. Smarter Pricing Becomes Part of the Product Experience

Let’s be honest: everyone loves a good deal.

But today’s shoppers aren’t simply looking for the lowest price. They’re becoming increasingly strategic about when they buy, where they compare, and whether the price they’re seeing actually represents good value.

In fact, Akeneo’s 2026 PX Pulse survey found that 79% of consumers have delayed a purchase because they expected the price to drop, while 77% have noticed the same product listed at different prices across retailers or platforms.

And with AI making it easier to compare products and prices in seconds, that scrutiny is only going to intensify.

This creates an interesting opportunity for retailers in 2027: bringing pricing intelligence and product information closer together.

Traditionally, pricing decisions and product content have often been managed separately. One team determines what a product costs, while another manages the specifications, descriptions, attributes, and experiences used to sell it.

But in an AI-driven shopping journey, those details are evaluated together.

5. Trust Becomes the Ultimate Competitive Advantage

Here’s a slightly uncomfortable truth about AI: just because it can generate an answer doesn’t mean that answer is correct.

As generative AI makes it easier than ever to produce product descriptions, images, reviews, and other content at scale, brands face a growing challenge distinguishing trustworthy information from inaccurate or misleading content.

We call this the AI slop trap: generating more content than ever before, without ensuring that the information behind it is actually reliable.

And consumers are paying attention.

DHL’s 2026 research found that privacy and trust concerns remain an important barrier to AI shopping adoption, with 48% of surveyed consumers expressing concerns in these areas.

In 2027, we expect first-party product data, verified specifications, trusted certifications, and transparent sourcing to become increasingly valuable.

After all, if an AI assistant is recommending a skincare product, safety accessory, or replacement automotive part, consumers need to know that the information it’s using is accurate.

And the importance of verified product information extends beyond AI.

Digital Product Passports (DPPs) are bringing new requirements for product transparency and traceability. Beginning February 18, 2027, battery passports become mandatory under EU rules for certain categories, including electric vehicle batteries, light means of transport batteries, and industrial batteries with capacities greater than 2 kWh. Requirements for additional product categories are expected to follow in later years.

For businesses, the opportunity is to build trust into their product experiences from the beginning, using accurate, governed information that can be verified, updated, and shared consistently.

Because in an increasingly AI-generated world, authenticity might become one of the most valuable things a brand can offer.

6. AI Moves Behind the Scenes, With Humans Still at the Helm

So far, we’ve focused on how AI will transform shopping for consumers.

But what about the people working behind the scenes to make all of those experiences possible?

In 2027, we expect one of the biggest shifts in eCommerce to happen inside the organizations themselves.

Instead of relying on individual AI tools to generate a product description or translate a few attributes, businesses will increasingly deploy specialized AI agents to support entire product information workflows.

It’s an exciting prospect, especially for product data teams that spend far too much time copying information between spreadsheets, fixing the same errors, and chasing missing attributes.

According to Deloitte’s 2026 Global Retail Industry Outlook, 68% of surveyed retailers expect to adopt agentic AI within the next 12 to 24 months.

But moving from an impressive AI demonstration to a reliable, production-ready workflow is a different challenge entirely.

What happens when an agent generates an inaccurate technical specification? Or publishes an unapproved sustainability claim? Or changes information that another agent is already working on?

That’s why AI governance will become an increasingly important operational discipline.

Businesses will need clear rules around what agents can access, modify, approve, and publish, as well as visibility into their actions and defined escalation paths when human judgment is needed.

And importantly, this doesn’t mean removing people from the process. It means changing what they’re responsible for.

Instead of manually completing every repetitive task, product experts can focus on guiding AI-powered workflows, reviewing exceptions, validating important decisions, and ensuring that automated systems remain aligned with business objectives.

7. Omnichannel Shopping Becomes More Connected Than Ever

With so much attention on AI assistants and autonomous purchasing, it would be easy to assume that traditional shopping channels are on their way out.

But the reality is more complicated.

Consumers aren’t abandoning physical stores, social media, marketplaces, or retailer websites. They’re simply adding more ways to discover and purchase products.

And they expect those experiences to work together.

DHL’s 2026 eCommerce Trends Report found that 63% of surveyed businesses sell through social channels, while 45% of shoppers buy through social commerce.

Meanwhile, physical stores continue to play an important role in product discovery, evaluation, and fulfillment.

The challenge for brands is making sure the experience remains consistent as customers move between channels.

A product promoted on social media needs to have the same accurate specifications as the one displayed on a marketplace. An AI recommendation should reflect current availability. A customer checking a product in-store should be able to access reliable information about sizes, materials, compatibility, and delivery options.

And all of those experiences need to be supported by fulfillment and service that live up to expectations.

The real opportunity in 2027 is creating a connected commerce ecosystem where every touchpoint tells the same trustworthy product story, whether the customer is browsing in person, scrolling through social media, or asking an AI agent to make a recommendation.

How to Prepare for 2027

If there’s one thing we’ve learned from watching eCommerce evolve over the past few years, it’s that the future rarely arrives in one neat, predictable package.

Some shoppers will enthusiastically hand their purchasing decisions over to AI agents. Others will still want to visit a store, speak to a knowledgeable employee, and inspect a product before making a decision.

Most will probably do a little bit of both.

But across all seven trends, one common thread emerges: the quality of the product information behind the experience matters more than ever.

AI agents can’t confidently recommend products they don’t understand. Personalized experiences can’t answer detailed questions without comprehensive product attributes. Dynamic pricing can’t deliver credible value when product and pricing information are disconnected. And autonomous workflows can’t scale responsibly without governance, traceability, and human oversight.

For brands, manufacturers, distributors, and retailers, preparing for 2027 starts with three priorities: velocity, comprehensiveness, and trust.

Product information needs to move quickly enough to keep pace with changing markets and channels. It needs to be comprehensive enough to answer increasingly specific customer questions. And it needs to be trustworthy enough for both humans and AI agents to rely on it.

At Akeneo, we believe the next generation of commerce will be powered by AI, but guided by human expertise and built on trusted product information.

After all, the future of shopping might involve fewer search bars, smarter AI agents, and buying journeys we haven’t even imagined yet.

But whether your next customer is a person, an algorithm, or some combination of the two, they’ll still need one thing before they’re ready to buy:

A good reason to trust your product, and that starts with the data behind it.

Autumn is officially here. The leaves are changing, the temperatures are dropping, and at Akeneo, we’re welcoming a new season of innovation.

But while the calendar might be turning, one thing isn’t slowing down: the pace of AI transformation.

Earlier this year, our Summer Release introduced expanded agentic capabilities through Ziggy, putting AI to work across the product information lifecycle. Now, with the Akeneo Autumn Release 2026, we’re taking the next step.

As AI takes on more responsibility, businesses need confidence that their product data is accurate, their processes are governed, and every change happens with the right oversight.

That’s exactly what this release delivers.

We’re introducing new ways to accelerate product data operations, strengthen governance, and help organizations deliver exceptional product experiences at enterprise scale.

Let’s take a look at what’s new.

Velocity: Moving Faster From Source to Storefront

Product data teams are under constant pressure to do more with less: onboard products faster, keep information accurate, and deliver consistent experiences across an expanding number of channels.

The Autumn Release introduces new capabilities that combine AI and automation to accelerate work throughout the product data lifecycle.

With enhanced Akeneo PX Insights, businesses gain greater visibility into how their products are discovered and understood by AI-powered search engines. Automated AI Discovery and AI Readiness analyses help identify missing product information, prioritize improvements, and turn insights into actionable opportunities. Working alongside Ziggy, teams can translate those insights into governed catalog improvements at scale.

New product model suggestions also help teams bring structure to growing catalogs by automatically identifying products that could be organized into shared models and hierarchies, reducing manual effort while maintaining oversight.

And with expanded Ziggy and Shopify integration, organizations can better understand synchronization issues and address underlying product data problems before they disrupt storefront experiences. The Akeneo App for Shopify now offers Standard and Enterprise editions, with Enterprise designed for businesses managing more complex, multi-store operations that require greater scale and more frequent synchronization.

Smarter Supplier Collaboration, Without the Back-and-Forth

Supplier onboarding can be one of the most time-consuming parts of managing product information, which is why the Autumn Release introduces expanded Akeneo Supplier Data Manager capabilities.

With new Supplier Collaboration capabilities, organizations can delegate more of the work of collecting, updating, and maintaining product information directly to the people who know those products best: their suppliers.

Rather than relying on internal teams to chase down missing specifications or manually manage updates, businesses can request targeted changes and empower suppliers to take an active role in maintaining accurate product data.

But delegating the work doesn’t mean giving up control. Internal teams can review supplier contributions before accepting changes into their trusted product catalog.

The result is less administrative work, faster collaboration, and greater confidence in the information flowing into the business.

Control: More Automation, Without Losing Oversight

As AI agents become more involved in product operations, an important question emerges: who’s in control of the changes they make?

That’s why every new AI capability in the Autumn Release operates within Akeneo’s AI harness, providing the permissions, business context, traceability, and human approval workflows necessary for AI to act reliably.

Agents operate within established permissions, their actions are traceable, and people retain oversight of proposed changes.

For organizations managing increasingly complex enterprise environments, that control extends beyond AI.

New Enterprise Access & Security Governance capabilities, available with Akeneo Product Cloud Premium, help businesses automate identity and access management through Just-in-Time (JIT) User Provisioning, SCIM, and SIEM integration.

These capabilities simplify user onboarding, keep access aligned with corporate identity policies, and strengthen security monitoring without creating more administrative work.

As the IT team at La Maison Simons, an Akeneo customer testing the new functionality, explains:

“Managing access for both our employees and external partners requires flexibility without compromising security. With Just-in-Time User Provisioning, we can support different login experiences, whether through SSO or direct login, while ensuring users have the right access from the start. It’s a valuable step toward simplifying user management while maintaining the security our business needs.”

Additional Supplier Enterprise Security capabilities bring SSO and multi-factor authentication to supplier collaboration, extending secure access beyond internal teams.

Governance: Defining What “Ready” Really Means

For years, product data teams have relied on completeness scores to determine whether products are ready for market.

But here’s the problem: a product can be 100% complete and still not be ready to sell.

A manufacturer might have every required field populated but be missing critical compatibility information. A beauty brand might have a complete description that hasn’t been approved for a particular market.

With new Product Readiness and Readiness Policies, Akeneo is helping businesses move beyond checking boxes.

Organizations can now define what “ready” means for specific business objectives, channels, markets, and locales by establishing rules around product attributes, values, descriptions, and relationships.

Teams can identify exactly what’s preventing a product from moving forward, prioritize improvements, and create more consistent processes across their organization.

And as governance requirements become more complex, Akeneo Product Cloud Premium provides enterprise-scale readiness capabilities, supporting thousands of policies across extensive catalogs, brands, markets, and business units.

Making Complex Product Relationships Simple

Governance isn’t just about individual products. It’s also about understanding how products relate to one another.

For manufacturers, distributors, and retailers managing complex catalogs, those relationships can quickly become difficult to maintain.

Product Compatibility, available with Akeneo Product Cloud Premium, brings dynamic product relationships directly into PIM.

Teams can define reusable rules connecting products, spare parts, accessories, and complementary items, with compatible relationships automatically maintained as catalogs evolve.

Instead of managing endless spreadsheets or manually updating associations, businesses can establish consistent, governed relationships that scale alongside their operations.

It’s another way Premium helps enterprise organizations manage greater product complexity without sacrificing agility.

Turning Over a New Leaf in Product Experience

Every new season brings an opportunity to rethink what’s possible.

And with the Akeneo Autumn Release 2026, we’re helping businesses do exactly that.

By connecting intelligent automation with product readiness, supplier collaboration, compatibility management, AI discoverability, and enterprise governance, Akeneo is making it easier to manage increasingly complex product ecosystems without multiplying manual work.

Ready to turn over a new leaf in your product data strategy?

Explore Akeneo’s latest release and discover how we’re helping businesses deliver smarter, more connected product experiences from source to sale.

Snowleader was not short on product knowledge. It was short on hours.

The outdoor retailer manages around 450,000 SKUs across more than 400 brands, with as many as 17,000 new products arriving during a seasonal launch. Writing every description from a blank page was never going to scale comfortably.

So the team introduced AI-generated product descriptions built from structured product data. Prompts were configured by product family and brand tone, while the content team remained responsible for reviewing and refining the results. 

According to Snowleader, the approach reduced the time spent writing descriptions by around 20%. That is a meaningful win: less repetitive work, faster launches, and more time for storytelling and optimization. 

It is also a good example of what a successful first AI use case looks like.

But it is not the end of the story.

Once one agent starts delivering value, the next questions arrive quickly. Could another agent onboard supplier files? Could one enrich missing attributes? Could a Quality Agent validate the result before the Content Agent acts? Could a Localization Agent prepare the approved copy for every market?

The pilot proves that AI can complete a task. What comes next determines whether it can become part of a product operation your teams can actually rely on.

The Gap Between Adoption and Impact

Businesses are not waiting to experiment with AI. McKinsey’s 2026 State of AI survey found that nearly nine in ten respondents reported regular AI use in at least one business function.

However, scaling is a different matter; the same survey found that only 44% said AI was being scaled across their enterprise, and just 37% attributed any EBIT impact to it. McKinsey also found that the highest-performing organizations were more likely to redesign workflows around AI rather than simply insert AI into existing processes.

That distinction matters for product data teams.

A pilot usually works with a carefully selected product family, a limited number of sources, and a well-understood workflow. Production introduces less tidy conditions: conflicting supplier values, unusual product structures, updated channel requirements, new markets, approval delays, and claims that require legal or technical judgment.

The model has not suddenly become less capable. The job has simply become real.

The Second Agent Changes the Conversation

One agent can often operate as a contained project, but as soon as you add another, the dependencies become more visible.

Imagine that your Supplier Data Agent updates a product specification. Should the Enrichment Agent act immediately, or wait for validation? Should the Content Agent regenerate the description? If the content changes, should every translated version be updated? Should the Syndication Agent pause publication until the new information is approved?

Without a shared operating foundation, agents may use different copies or versions of the same product information. The Content Agent may act before the Quality Agent finishes its review. The Localization Agent may translate content that is still changing upstream. Each agent can appear successful in isolation while the full process becomes harder to control.

What Orchestrating Multiple AI Agents Requires

Moving beyond the pilot requires more than connecting an agent to additional data. Four capabilities become especially important:

One shared product record: Every agent needs access to the same governed source of product truth. If attributes conflict or ownership is unclear, the agent may confidently act on the wrong value.

Clear sequencing: Supplier ingestion, enrichment, validation, content generation, localization, and activation are connected stages. An agent needs to know not only what to do, but when it is safe to do it.

Shared business context: A generic model does not know your taxonomy, approved terminology, brand voice, product-family rules, or marketplace requirements. That knowledge has to be structured so it can be reused across agents rather than recreated inside every prompt.

Oversight and traceability: People still need to review decisions involving uncertainty, compliance, technical accuracy, or brand judgment. Teams must also be able to see which source an agent used, what it recommended, who approved it, and what ultimately changed.

This lifecycle approach reflects broader guidance such as the NIST AI Risk Management Framework, which encourages organizations to incorporate trustworthiness and risk management throughout the design, deployment, use, and evaluation of AI systems, not only at the pilot stage.

What to Do After a Successful Pilot

Resist the temptation to move directly from a polished demonstration to the entire catalog.

Start by testing the agent against a representative mix of product families, suppliers, exceptions, and channels. Confirm that it can retrieve the right context, identify authoritative sources, and handle imperfect inputs.

Next, define the decision boundaries. Which actions can happen automatically? Which require review? Which must remain human-led? Make sure reviewers receive the evidence behind a recommendation, not only the final answer.

Finally, measure the process rather than the novelty. Compare manual effort, cycle time, accuracy, acceptance rates, review time, publication success, and downstream impact before and after deployment.

A pilot should lead to one of four decisions: expand, refine, redesign, or stop.

Your first agent shows what AI can do. The choices you make after the pilot determine whether the next agent becomes another isolated experiment, or part of a connected product operation where every new capability builds on the data, governance, and lessons already in place.

It is the same reflex every November: 30% off the entire site. The revenue chart looks impressive by Cyber Monday. Then January arrives, and the margin report tells a different story.

The real question is not how much to discount. It is which price, for which product, at which moment of its lifecycle. Answering it requires something most retailers still lack on Black Friday: product data and pricing working from the same page.

I. The Blanket Discount Trap

A blanket discount treats every product the same way. Your shoppers do not. According to BCG (2025), about 77% of consumers say they sometimes or often delay purchases earlier in the year to take advantage of year-end sales. Part of your Black Friday volume is demand you would have captured at full price, just a few weeks earlier.

The damage adds up fast. Margin disappears on products that never needed a discount. Prices drop to match a competitor, even when the products are not truly comparable.

Underneath sits a structural problem. The product lives in the PIM. The price lives in a spreadsheet. Black Friday is the moment that disconnect gets stress-tested, at scale and under time pressure.

If your pricing team cannot see what makes a product different, how can it decide what that product is worth?

II. Stage 1, Before the Rush: Lock In a Solid Everyday Price

Black Friday starts in October. In the EU, any advertised discount must be based on the lowest price applied in the previous 30 days. The European Commission’s 2026 sweep found that 30% of the 314 online traders checked presented discounts incorrectly. Your October prices set the reference for your November promotions.

Akeneo PricingHUB handles these price regulations natively: your teams set each constraint once as a business rule, and it applies to every price and promotion. Compliance stops being a last-minute check.

Trust is fragile too. Simon-Kucher’s 2025 Holiday Shopping Report found that only 46% of US shoppers believe advertised discounts fully reflect true savings. A credible discount starts with a credible everyday price.

Keep your price ladder intact

Picture a grocery or DIY retailer with two rules: keep a 20% margin on organic products, and price its private label 15% below the equivalent national brand. Without structured data, the pricing team tags thousands of SKUs by hand before it can set a single rule. When attributes like “organic” or “private label” already live in the PIM, the rule is set once and applies across the whole catalog. Your price ladder survives the promotion.

Get your range right before you discount it

Now take a running shoe sold in 15 sizes, 3 colors and two versions, Standard and Pro. Size 42 should never cost more than size 43. The Pro should always sit above the Standard. With a clean product structure, these consistency rules apply automatically, so your discounts start from a coherent range rather than a patchwork.

III. Stage 2, The Promotion: Discount Where It Pays

Every discount should earn its place. BCG (2025) recommends targeted promo plans that guide consumers toward assortments where the retailer has surplus or margin advantage. In practice: discount where stock is high or margin can absorb it, and hold your price everywhere else.

Pick the products shoppers actually compare

Price elasticity measures how much a price change shifts what people buy. On Black Friday, a TV is highly elastic: shoppers compare it across retailers, and a small price gap moves volume. The HDMI cable sold with it is not. Few people compare cable prices, so a discount there gives away margin for almost no extra sales.

Products also influence each other. Discount one TV, and part of its extra volume comes from a similar model you still sell at full price. This transfer, called cannibalization, is easier to anticipate when the system knows which products are true substitutes. That knowledge comes from product attributes: screen size, technology, brand, range.

Launch in hours, not weeks

Picture a seasonal campaign where merchandising, product marketing and pricing each work from their own files. Reconciling product lists with promotional prices can take up to 14 days of manual alignment. When product and pricing teams work from the same data, there is nothing left to extract or reconcile. The same campaign can go live in hours.

The right price on the right channel

Your stores, your website, your country sites and your marketplaces do not face the same competition. Each deserves its own price. When optimized prices flow back into the PIM, they reach every channel alongside the right product content, with no manual exports.

IV. Stage 3, Under Competitive Pressure: Defend Your Premium

It is Black Friday week, and your best-selling white t-shirt is priced at €15. A competitor drops its white t-shirt to €5. A disconnected pricing tool sees only the gap and matches it, wiping out €10 of margin per unit.

Now add product context. Your t-shirt carries the “organic cotton” attribute. Instead of slashing the price, your team enriches the product page to highlight that quality and holds the €15. The difference between destroying margin and protecting it is product context.

V. Stage 4, After the Peak: Clear Stock Without Burning Margin

Black Friday ends, but the stock does not disappear. Some products leave the week with more inventory than planned, often in specific stores or regions. A blanket markdown before Christmas would repeat the same mistake twice.

A planned markdown starts at category level, not SKU by SKU. It accounts for the stock left on each product, in each location. It also measures how a price cut on one product shifts demand across the whole category, so a markdown does not simply pull sales away from similar products still at full price. The result is a discount path that clears remaining stock before the winter sales while protecting as much margin as possible.

Products that sell well keep their price. Slow movers get the discount they actually need.

VI. AI Recommends, Your Teams Decide

None of this means handing Black Friday to a black box. Automation does not mean losing control. The AI surfaces patterns and recommends prices. Pricing experts validate them and turn them into lasting business rules.

Black Friday is also a workload peak. Thousands of prices to check, adjust and push to every channel, often by hand and against the clock. Akeneo PricingHUB automates that execution: one of our European retail customers now rolls out its optimized prices more than 50 times faster.

Automation does not mean handing Black Friday to a black box. The AI recommends prices and explains every one of them. Your teams validate, adjust and keep the final say, supported by a dedicated team of pricing experts.

Conclusion

Black Friday rewards retailers who know exactly what they sell and what it is worth. When product data and pricing are connected, every discount becomes a deliberate choice: which product, which channel, which moment. That is how Black Friday becomes a controlled decision rather than a bet on your margin.

About Akeneo PricingHUB

Our platform is designed to optimise prices throughout the product lifecycle for retailers and e-commerce businesses. Our objective? To optimise your prices in a simple, automated and data-driven way. Our intuitive solution is centred around AI, a consumer-centric approach and pricing expertise.

👉 Meet our pricing expert here 

The most revealing moment in an AI project often arrives five minutes after the demo ends.

The product description provided looks polished, or the product recommended looks accurate and relevant.

But then someone asks: Which source did it use? Is that claim approved in every market? Does it have the current price? Who reviews the change before it reaches a live channel?

Suddenly, the future of commerce looks less like science fiction and more like a missing attribute in a spreadsheet.

That tension sits at the heart of IT departments across the globe today, and what we set out to learn more about in our annual survey of IT and technology decision-makers. 

On paper, confidence is high: 95% of IT leaders say their organization is ready for AI-driven commerce, and 87% expect AI budgets to increase. 

But here’s the catch: 89% also spend more than a quarter of every AI project preparing data. 

AI Is Exposing Infrastructure Problems It Did Not Create

Nearly two-thirds of IT leaders say their department now owns AI strategy, meaning they’re the ones responsible for making AI work inside the systems the business already relies on: ERP, PIM, DAM, supplier feeds, commerce platforms, marketplaces, and regional workflows.

A team may be able to vibe-code an impressive prototype over a weekend, but it cannot vibe-code trusted product information, resilient integrations, permissions, approvals, and audit trails across brands and markets.

AI did not create brittle integrations, local taxonomies, manual mapping, or disconnected workflows, however it is making their cost harder to ignore. Every new use case inherits that technical debt unless the underlying product information foundation is modernized.

“AI-Ready” Data Still Requires a Great Deal of Human Work

While 95% of IT leaders say they are ready for AI commerce and 92% rate their data quality and structure as good or excellent, 53% spend between 26% and 50% of an AI project preparing data. Another 36% spend more than half.

That preparation means chasing missing attributes, reconciling duplicate SKUs, normalizing units and taxonomies, checking regional prices and compliance fields, resolving conflicting values, and adapting information for different channels.

It is repetitive, time-consuming work the wider business rarely sees. When it is done well, nothing dramatic happens; the AI simply produces an answer that is not wrong. Yet that invisible effort is often what holds the project together.

A Policy Is Not the Same as Operational Governance

Almost every organization surveyed (99%) either has an AI governance policy or is developing one. But a policy cannot always answer the questions that matter when AI touches live product information:

Who changed the record? Which model made the recommendation? What data did it use? Who approved the output? Where was the change published?

That traceability, sometimes called decision lineage, matters when a change involves a price, warranty, compliance claim, promotion, or product description. Governance becomes useful when it is built into workflows through permissions, human review, monitoring, and audit trails, not when it exists only on a board slide.

Automation Solves One Scale Problem, But Can Create Another

In our survey, we found that IT automation and IT operations outrank customer-facing use cases such as conversational AI and personalized search as investment priorities. That makes sense: managing millions of attributes, products, channels, and markets has already exceeded what many teams can handle manually.

The complication is that every narrowly focused agent added to solve one bottleneck becomes another component to coordinate. Without a shared foundation, multiple agents may enrich, translate, validate, and activate the same records without understanding what the others have changed.

This is where an AI harness becomes useful. It is everything around the model (permissions, memory, guardrails, approvals, and orchestration) that allows it to act reliably inside business systems. 

Product Discoverability Is Becoming a Data Problem

Seventy-nine percent of organizations are investing in or piloting Answer Engine Optimization (sometimes called Generative Engine Optimization) as product discovery shifts toward AI.

But appearing in an AI-generated recommendation is not simply a new version of ranking for a keyword. An assistant must understand the product, compare it with alternatives, verify its attributes and claims, and trust that the information is current.

Incomplete titles, missing identifiers, inconsistent specifications, unsupported claims, or stale availability can keep a product out of the consideration set entirely.

The Real Constraint Is the Foundation

Enthusiasm for AI is not fading. Budgets are growing, ownership is becoming clearer, and organizations are actively investing.

What is holding AI back is the gap between a promising model and the operational environment required to trust it. More funding will not remove the data-preparation tax when every new project begins with the same incomplete attributes, inconsistent taxonomies, regional variations, and fragile integrations.

The next meaningful AI milestone may not be a dazzling demo. It may be a far less cinematic moment: an agent updates the correct product record, follows the right rules, routes the change to a human, and publishes accurate information to the right channel.

That is the moment AI stops being an experiment and starts becoming infrastructure.

Download the full report, The Agentic Commerce Reality Check, to explore all 10 hard truths from IT leaders and assess whether your product information foundation is ready for agentic commerce.

It has been a big year at Akeneo.

We’ve expanded what Akeneo Product Cloud can do. We’ve brought new technologies and expertise into the company. We’ve introduced agentic AI capabilities designed to change how product teams work. 

We’ve continued accelerating our presence in the U.S, and we’ve worked with brands including LG Electronics USA, Tiffany & Co., Sweaty Betty, and more to transform their product operations.

Now, there’s another recognition to share.

Akeneo has been recognized as a Leader in the first-ever 2026 Gartner® Magic Quadrant™ for Product Information Management Solutions.

Gartner evaluates providers based on Ability to Execute and Completeness of Vision. For us, being positioned as a Leader in the inaugural Gartner® Magic Quadrant™ for Product Information Management Solutions is an important moment to us, but it is not one we see as ours alone.

We feel it reflects years of collaboration with customers who challenge us to solve increasingly complex product-information problems, partners who extend and strengthen the Akeneo ecosystem, and teams across Akeneo who keep pushing the Product Cloud forward.

A Recognition Built With our Customers and Partners

The first Gartner® Magic Quadrant™ for Product Information Management Solutions marks a significant moment for the PIM category itself, but for Akeneo, it also gives us an opportunity to recognize the ecosystem that has helped shape where we are today.

Our customers bring the real-world complexity: global catalogs, intricate data models, thousands of attributes, multiple markets and channels, changing regulations, and increasingly AI-driven commerce experiences.

Our partners bring the implementation expertise, integrations, services, and market knowledge that help turn technology into real business outcomes.

And that collaboration shows up in the feedback we receive. Akeneo has a 4.7 out of 5 rating based on 140 customer reviews on Gartner Peer Insights™.

One Engineering Manager in the retail industry shared:

“Akeneo’s flexibility in modeling complex product data is a major strength. The attributes and family structure make it easy to scale and manage large catalogs.”

And a Director of Operations in manufacturing noted:

“From the start, the team engaged in demonstrating Akeneo’s features and how it aligned with our needs and solved our pain points we experienced with the previous platform.”

Their skills, feedback, ambition, and expectations continually shape what we build next.

So, to our customers and partners: thank you. Thank you for challenging us, collaborating with us, and helping us keep raising the bar for what product information can do. We see this milestone as one we share with you.

From AI Assistance to Agentic Action

Our Summer Release introduced Agentic Ziggy, an agentic layer embedded within Akeneo Product Cloud to help teams move from simply managing product information to acting on it.

Customers are already using Ziggy to surface insights, simplify complex updates, and automate time-consuming work. At tectake, for example, Ziggy generated 200 automation rules in minutes, which was  work that would previously have taken around two hours manually.

Growing Momentum Across the Market

That product momentum is being matched by business momentum.

Akeneo continues to expand across key markets, including the U.S., as more organizations look to product information as a foundation for faster operations, stronger digital experiences, and emerging AI-powered workflows. That growth also creates a valuable feedback loop, helping us sharpen the roadmap and build for the real-world complexity our customers face.

And We’re Not Finished Yet – Stay Tuned!

Recognition in the first-ever Gartner® Magic Quadrant™ for Product Information Management Solutions comes at an exciting point in Akeneo’s journey.

This year has already brought strategic acquisitions, new customers, expanded capabilities, and an entirely new agentic layer within Akeneo Product Cloud.

And there is more coming.

Akeneo’s Autumn Release is coming soon, with additional functionality designed to expand what teams can do with their product information and how quickly they can turn it into action.

We’ll have much more to share shortly.

For now, we’re proud to see Akeneo positioned as a Leader in the first-ever Gartner® Magic Quadrant™ for Product Information Management Solutions, and grateful to the customers, partners, and teams who have helped us get here.

Download the 2026 Gartner® Magic Quadrant™ for Product Information Management Solutions and explore Akeneo’s position as a Leader.

——

Gartner, Magic Quadrant™ for Product Information Management Solutions, Andrew Davies, Eric Ahlm, Angel Berrios, Darren Livingstone, 12 October 2026.

Gartner, Peer Insights and Magic Quadrant are trademarks of Gartner, Inc. and/or its affiliates.

Gartner Peer Insights content consists of the opinions of individual end users based on their own experiences with the vendors listed on the platform, should not be construed as statements of fact, nor do they represent the views of Gartner or its affiliates. Gartner does not endorse any vendor, product or service depicted in this content nor makes any warranties, expressed or implied, with respect to this content, about its accuracy or completeness, including any warranties of merchantability or fitness for a particular purpose.

Gartner does not endorse any company, vendor, product or service depicted in its publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner publications consist of the opinions of Gartner’s business and technology insights organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this publication, including any warranties of merchantability or fitness for a particular purpose.